1. Field of the Invention
The present invention relates to a method and apparatus for stabilizing output beam parameters of a gas discharge laser. More particularly, the present invention relates to maintaining an optimal gas mixture composition over long, continuous operating or static periods using very small gas injections.
2. Discussion of the Related Art
Pulsed gas discharge lasers such as excimer and molecular lasers emitting in the deep ultraviolet (DUV) or vacuum ultraviolet (VUV) have become very important for industrial applications such as photolithography. Such lasers generally include a discharge chamber containing two or more gases such as a halogen and one or two rare gases. KrF (248 nm), ArF (193 nm), XeF (350 nm), KrCl (222 nm), XeCl (308 nm), and F2 (157 nm) lasers are examples.
The efficiencies of excitation of the gas mixtures and various parameters of the output beams of these lasers vary sensitively with the compositions of their gas mixtures. An optimal gas mixture composition for a KrF laser has preferred gas mixture component ratios around 0.1% F2/1% Kr/98.9% Ne (see U.S. Pat. No. 4,393,505, which is assigned to the same assignee and is hereby incorporated by reference). A F2 laser may have a gas component ratio around 0.1% F2/99.9% Ne or He or a combination thereof (see U.S. Pat. No. 6,157,662, which is assigned to the same assignee and is hereby incorporated by reference). Small amounts of Xe may be added to rare gas halide gas mixtures, as well (see U.S. patent application Ser. No. 09/513,025, which is assigned to the same assignee and is hereby incorporated by reference; see also R. S. Taylor and K. E. Leopold, Transmission Properties of Spark Preionization Radiation in Rare-Gas Halide Laser Gas Mixes, IEEE Journal of Quantum Electronics, pp. 2195-2207, vol. 31, no. 12 (December 1995). Any deviation from the optimum gas compositions of these or other excimer or molecular lasers would typically result in instabilities or reductions from optimal of one or more output beam parameters such as beam energy, energy stability, temporal pulse width, temporal coherence, spatial coherence, discharge width, bandwidth, and long and short axial beam profiles and divergences.
Especially important in this regard is the concentration (or partial pressure) of the halogen, e.g., F2, in the gas mixture. The depletion of the rare gases, e.g., Kr and Ne for a KrF laser, is low in comparison to that for the F2. FIG. 1 shows laser output efficiency versus fluorine concentration for a KrF laser, showing a decreasing output efficiency away from a central maximum. FIG. 2 shows how the temporal pulse width (pulse length or duration) of KrF laser pulses decrease with increasing F2 concentration. FIGS. 3-4 show the dependence of output energy on driving voltage (i.e., of the discharge circuit) for various F2 concentrations of a F2 laser. It is observed from FIGS. 3-4 that for any given driving voltage, the pulse energy decreases with decreasing F2 concentration. In FIG. 3, for example, at 1.9 kV, the pulse energies are around 13 mJ, 11 mJ and 10 mJ for F2 partial pressures of 3.46 mbar, 3.16 mbar and 2.86 mbar, respectively. The legend in FIG. 3 indicates the partial pressures of two premixes, i.e., premix A and premix B, that are filled into the discharge chamber of a KrF laser. Premix A comprised substantially 1% F2 and 99% Ne, and premix B comprised substantially 1% Kr and 99% Ne. Therefore, for the graph indicated by triangular data points, a partial pressure of 3.46 mbar for premix A indicates that the gas mixture had substantially 3.46 mbar of F2 and a partial pressure of 3200 mbar for premix B indicates that the gas mixture had substantially 32 mbar of Kr, the remainder of the gas mixture being the buffer gas Ne. FIG. 5 shows a steadily increasing bandwidth of a KrF laser with increasing F2 concentration.
In industrial applications, it is advantageous to have an excimer or molecular fluorine laser capable of operating continuously for long periods of time, i.e., having minimal downtime. It is desired to have an excimer or molecular laser capable of running non-stop year round, or at least having a minimal number and duration of down time periods for scheduled maintenance, while maintaining constant output beam parameters. Uptimes of, e.g., greater than 98% require precise control and stabilization of output beam parameters, which in turn require precise control of the composition of the gas mixture.
Unfortunately, gas contamination occurs during operation of excimer and molecular fluorine lasers due to the aggressive nature of the fluorine or chlorine in the gas mixture. The halogen gas is highly reactive and its concentration in the gas mixture decreases as it reacts, leaving traces of contaminants. The halogen gas reacts with materials of the discharge chamber or tube as well as with other gases in the mixture. Moreover, the reactions take place and the gas mixture degrades whether the laser is operating (discharging) or not. The passive gas lifetime is about one week for a typical KrF-laser.
During operation of a KrF-excimer laser, such contaminants as HF, CF4, COF2, SiF4 have been observed to increase in concentration rapidly (see G. M. Jurisch et al., Gas Contaminant Effects in Discharge-Excited Krf Lasers, Applied Optics, Vol. 31, No. 12, pp. 1975-1981 (Apr. 20, 1992)). For a static KrF laser gas mixture, i.e., with no discharge running, increases in the concentrations of HF, O2, CO2 and SiF4 have been observed (see Jurisch et al., above).
One way to effectively reduce this gas degradation is by reducing or eliminating contamination sources within the laser discharge chamber. With this in mind, an all metal, ceramic laser tube has been disclosed (see D. Basting et al., Laserrohr für halogenhaltige Gasentladungslaser” G 295 20 280.1, Jan. 25, 1995/Apr. 18, 1996 (disclosing the Lambda Physik Novatube, and hereby incorporated by reference into the present application)). FIG. 6 qualitatively illustrates how using a tube comprising materials that are more resistant to halogen erosion (plot B) can slow the reduction of F2 concentration in the gas mixture compared to using a tube which is not resistant to halogen erosion (plot A). The F2 concentration is shown in plot A to decrease to about 60% of its initial value after about 70 million pulses, whereas the F2 concentration is shown in plot B to decrease only to about 80% of its initial value after the same number of pulses. Gas purification systems, such as cryogenic gas filters (see U.S. Pat. Nos. 4,534,034, 5,136,605, 5,430,752, 5,111,473 and 5,001,721 assigned to the same assignee, and hereby incorporated by reference) or electrostatic particle filters (see U.S. Pat. No. 4,534,034, assigned to the same assignee and 5,586,134, each of which is incorporated by reference) are also being used to extend KrF laser gas lifetimes to 100 million shots before a new fill is advisable.
It is not easy to directly measure the halogen concentration within the laser tube for making rapid online adjustments (see U.S. Pat. No. 5,149,659 (disclosing monitoring chemical reactions in the gas mixture)). Therefore, it is recognized in the present invention that an advantageous method applicable to industrial laser systems includes using a known relationship between F2 concentration and a laser parameter, such as one of the F2 concentration dependent output beam parameters mentioned above. In such a method, precise values of the parameter would be directly measured, and the F2 concentration would be calculated from those values. In this way, the F2 concentration may be indirectly monitored.
Methods have been disclosed for indirectly monitoring halogen depletion in a narrow band excimer laser by monitoring beam profile (see U.S. Pat. No. 5,642,374, hereby incorporated by reference) and spectral (band) width (see U.S. Pat. No. 5,450,436, hereby incorporated by reference). Neither of these methods is particularly reliable, however, since beam profile and bandwidth are each influenced by various other operation conditions such as repetition rate, tuning accuracy, thermal conditions and aging of the laser tube. That is, the same bandwidth can be generated by different gas compositions depending on these other operating conditions.
An advantageous technique monitors amplified spontaneous emission (ASE), and is described in U.S. Pat. No. 6,243,406 (assigned to the same assignee and hereby incorporated by reference). The ASE is very sensitive to changes in fluorine concentration, and thus the fluorine concentration may be monitored indirectly by monitoring the ASE, notwithstanding whether other parameters are changing and effecting each other as the fluorine concentration in the gas mixture changes.
It is known to compensate the degradation in laser efficiency due to halogen depletion by steadily increasing the driving voltage of the discharge circuit to maintain the output beam at constant energy. To illustrate this, FIG. 7 shows how at constant driving voltage, the energy of output laser pulses decreases with pulse count. FIG. 8 then shows how the driving voltage may be steadily increased to compensate the halogen depletion and thereby produce output pulses of constant energy.
One drawback of this approach is that output beam parameters other than energy such as those discussed above with respect to FIGS. 1-5 affected by the gas mixture degradation will not be correspondingly corrected by steadily increasing the driving voltage. FIGS. 9-11 illustrate this point showing the driving voltage dependencies, respectively, of the long and short axis beam profiles, short axis beam divergence and energy stability sigma. Moreover, at some point the halogen becomes so depleted that the driving voltage reaches its maximum value and the pulse energy cannot be maintained without refreshing the gas mixture.
It is desired to have a method of stabilizing all of the output parameters affected by halogen depletion and not just the energy of output pulses. It is recognized in the present invention that this is most advantageously achieved by adjusting the halogen and rare gas concentrations themselves.
There are techniques available for replenishing a gas mixture by injecting additional rare and halogen gases into the discharge chamber between new gas fills and to methods including readjusting the gas pressure, e.g., by releasing gases from the laser tube (see especially U.S. Pat. Nos. 6,490,307 and 6,212,214, and also U.S. Pat. No. 6,243,406; and U.S. Pat. Nos. 5,396,514 and 4,977,573, each of which is assigned to the same assignee and hereby incorporated by reference). A more complex system monitors gas mixture degradation and readjusts the gas mixture using selective replenishment algorithms for each gas of the gas mixture (see U.S. Pat. No. 5,440,578, hereby incorporated by reference). One technique uses an expert system including a database of information and graphs corresponding to different gas mixtures and laser operating conditions (see the '214 patent, mentioned just above). A data set of driving voltage versus output pulse energy, e.g., is measured and compared to a stored “master” data set corresponding to an optimal gas composition such as may be present in the discharge chamber after a new fill. From a comparison of values of the data sets and/or the slopes of graphs generated from the data sets, a present gas mixture status and appropriate gas replenishment procedures, if any, may be determined and undertaken to reoptimize the gas mixture. Early gas replenishment procedures are described in the '573 patent (mentioned above).
Most conventional techniques generally produce some disturbances in laser operation conditions when the gas is replenished. For example, strong pronounced jumps of the driving voltage are produced as a result of macro-halogen injections (macro-HI) as illustrated in FIG. 12 (macro-HI are distinguished from micro-halogen injections, or μHI, as described in the '307 patent). The result of a macro-HI is a strong distortion of meaningful output beam parameters such as the pulse-to-pulse stability. For this reason, in some techniques, the laser is typically shut down and restarted for gas replenishment, remarkably reducing laser uptime (see U.S. Pat. No. 5,450,436).
The '307 patent referred to above provides a technique wherein gas replenishment is performed for maintaining constant gas mixture conditions without disturbing significant output beam parameters. The '307 patent describes a gas discharge laser system which has a discharge chamber containing a gas mixture including a constituent halogen-containing species, a pair of electrodes connected to a power supply circuit including a driving voltage for energizing the first gas mixture, and a resonator surrounding the discharge chamber for generating a laser beam.
A gas supply unit is connected to the discharge chamber for replenishing the gas mixture including the constituent halogen-containing species. The gas supply unit includes a gas inlet port having a valve for permitting a small amount of gas to inject into the discharge chamber to mix with the gas mixture therein. A processor monitors a parameter indicative of the partial pressure of the first constituent gas and controls the valve at successive predetermined intervals to compensate a degradation of the constituent halogen-containing species in the gas mixture.
The partial pressure of the halogen containing-species in the gas mixture is increased by an amount preferably less than 0.2 mbar, as a result of each successive injection. The gaseous composition of the injected gas is preferably 1%-5% of the halogen-containing gas and 95%99% buffer gas, so that the overall pressure in the discharge chamber increases by less than 20 mbar, and preferably less than 10 mbar per gas injection.
The processor monitors the parameter indicative of the partial pressure of the halogen-containing gas and the parameter varies with a known correspondence to the partial pressure of the halogen gas. The small gas injections each produce only small variations in partial pressure of the halogen gas in the gas mixture of the laser tube, and thus discontinuities in laser output beam parameters are reduced or altogether avoided.
The constituent gas is typically a halogen containing molecular species such as molecular fluorine or hydrogen chloride. The constituent gas to be replenished using the method of the '307 patent may alternatively be an active rare gas or gas additive. The monitored parameter may be any of time, shot count, driving voltage for maintaining a constant laser beam output energy, pulse shape, pulse duration, pulse stability, beam profile, bandwidth of the laser beam, energy stability, temporal pulse width, temporal coherence, spatial coherence, amplified spontaneous emission (ASE), discharge width, and long and short axial beam profiles and divergences, or a combination thereof. Each of these parameters varies with a known correspondence to the partial pressure of the halogen, and then halogen partial pressure is then precisely controlled using the small gas injections to provide stable output beam parameters.
The gas supply unit of the '307 patent preferably includes a small gas reservoir for storing the constituent gas or second gas mixture prior to being injected into the discharge chamber (see U.S. Pat. No. 5,396,514, which is assigned to the same assignee and is hereby incorporated by reference, for a general description of how such a gas reservoir may be used). The reservoir may be the volume of the valve assembly or an additional accumulator. The accumulator is advantageous for controlling the amount of the gas to be injected. The pressure and volume of the gases to be injected are selected so that the overall pressure in the discharge chamber will increase by a predetermined amount preferably less than 10 mbar, and preferably between 0.1 and 2 mbar, with each injection. As above, the halogen partial pressure preferably increases by less than 0.2 mbar and preferably far less such as around 0.02 mbar per injection. These preferred partial pressures may be varied depending on the percentage concentration of the halogen containing species in the gas premixture to be injected.
Injections may be continuously performed during operation of the laser in selected amounts and at selected small intervals. Alternatively, a series of injections may be performed at small intervals followed by periods wherein no injections are performed. The series of injections followed by the latent period would then be repeated at predetermined larger intervals. A comprehensive algorithm is desired for performing gas actions in order to better stabilize the gas composition in the laser tube, and correspondingly better stabilize significant parameters of the output beam of the excimer or molecular fluorine laser system.